Fig. 2

Simplified representation of the analytical method used in the experiment and analytical output. The images are divided into two folders based on the results of the histopathological analysis. Thereafter, the dataset is divided into a training, a validation and a test set. The training and the validation sets are used for the transfer-learning procedure with GoogleNet. A schematic and simplified representation of the output of the first convolutional layers is reported. Please note that the features represented become more complex during convolutions. Lastly, the retrained GoogleNet convolutional deep neural network is used to predict the labels for the test set. A confusion matrix is generated as a final output. n = number of images